How Do We Know if a Charter School is Really Succeeding? – Various Approaches to Investigating School Effectiveness October 2012 Missouri Charter Public Schools Association 2012 Conference Presented By: Otoniel Lopez Jing Zhu

Introduction Why data and evaluation? Evaluation components Impact/outcome studies Implementation studies Choosing the right design 1 Presentation Outline

Focus areas: K-12 Education, higher education, children and family services, youth development, juvenile justice, etc. Conducted six evaluations of two charter school programs spanning six years 2 Metis Associates National applied research and consulting firm (New York City, Philadelphia, Atlanta) Over 35 years of expertise in research/evaluation, grants development, and information technology

3 Why data and evaluation? Demonstrate program impact Identify successful practices and challenges Assess overall program fidelity Engage key stakeholders Facilitate the daily management of the grant Inform programmatic decisions Fulfill federal and state reporting requirements

4 Evaluation Components PurposeData sources and methods Impact/ Outcome Study Assess program impact on: 1.Academic performance 2.Customer impact and satisfaction Statistical analyses Review of school characteristics and their association with outcomes Stakeholder surveys Analysis of demographic, program participation, academic achievement and attendance data Implementation Study Assess implementation regarding: 1.Program fidelity 2.Promising practices, challenges and lessons learned Review of project documentation Interviews with project staff and partners Observations of cross-school activities

Randomized controlled trial (RCT) –The gold standard –Random assignment of students, classes or schools –A number of long-standing concerns (e.g., ethical, logistical, and financial) –Attrition and other issues 5 Impact Study Designs (I)

Quasi-experimental design (QED) –Need for a comparison group Naturally occurring Statistically well-matched –Common matching characteristics (baseline achievement, gender, race/ethnicity, ELL status, poverty status, etc.) –Assess baseline equivalence of two groups –Cannot control for potential unobserved differences between groups 6 Impact Study Designs (II)

The What Works Clearinghouse (WWC) –Initiative of the U.S. DOE’s IES –Started in 2002, reports since 2005 Three possible study ratings –Meets WWC Evidence Standards without Reservations (RCT with low attrition) –Meets WWC Evidence Standards with Reservations (RCT with high attrition OR QED; must establish baseline equivalence) –Does Not Meet WWC Evidence Standards 7 WWC Study Ratings

RCT studies –Random lottery (oversubscription to enrollment) Gleason, P., Clark, M., Tuttle, C. C., & Dwoyer, E. (2010). The Evaluation of Charter School Impacts: Final Report. Dobbie, W., & Fryer, R. G., Jr., (2009). Are High-Quality Schools Enough to Close the Achievement Gap? Evidence from a Social Experiment in Harlem. QED studies –Statistical matching of students Center for Research on Education Outcomes. (2011). Charter School Performance in Indiana. Center for Research on Education Outcomes. (2010). Charter School Performance in New York City. 8 Rigorous Charter School Evaluations

Survey Research Observations CAPTURE experience of participants PROVIDE quantifiable data that can be used in associating that experience with other hard data (e.g., student achievement) MEASURE changes in perceptions overtime ASSESS instruction using quantitative tools developed from a set of standards or known best practices QUANTIFY a set of items or behaviors within a school or classroom Two Popular Qualitative Methods in Impact Studies for Charter Schools 9

Assessment of program fidelity Question of resources and capacity Are intended populations being reached? Are services appropriate? Alignment of outcomes and implementation –Logic model 10 Implementation Studies

11 Logic Model

Interviews with key personnel Focus Groups with a set of individuals closely tied to the particular program (e.g., teachers) Observations of instruction, faculty meetings, or school walkthroughs Some survey research Collection of program documentation 12 Methods for Collecting Implementation Data

Implementation: Provides ongoing data (i.e., formative) Provides a real-world look at what is actually going on at a school Does not require long periods to gather useful information Doesn’t require a comparison group Outcome: Measures program impact Can provide an evidence base Provides useful information to policy makers 13 Advantages of Implementation and Outcome Components

Complex designs vs. point-in-time descriptive studies Balancing design approaches in current economic climate Before identifying right fit: –Use of theories of change, logic models, information systems and self-evaluation to inform research. 14 One Size Doesn’t Fit All

What Works Clearinghouse (WWC) Official Website ( American Evaluation Association Online Resources ( American Education Evaluation Association ( Kellogg Foundation ( and-resources.aspx) and-resources.aspx W.K. Kellogg Foundation Logic Model Development Guide W.K. Kellogg Foundation Evaluation Handbook The Evaluator’s Institute ( Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2004).Evaluation: A Systematic Approach. (7 ed.). Thousand Oaks: Sage Publications, Inc. 15 Evaluation Resources

16 Otoniel Lopez Jing Zhu